Guidelines for the validation of machine learning predictions of species interactions

1. The prediction of species interactions is gaining momentum as a way to circumvent limitations in data volume. Yet, ecological networks are challenging to predict because they are typically small and sparse. Dealing with extreme class imbalance is a challenge for most binary classifiers, and there are currently no guidelines as to how predictive models can be trained for this specific problem.2. Using simple mathematical arguments and numerical experiments in which a variety of classifiers (for supervised learning) are trained on simulated networks, we develop a series of guidelines related to the choice of measures to use for model selection, and the degree of unbiasing to apply to the training dataset.3. Neither classifier accuracy nor the ROC-AUC are informative measures for the performance of interaction prediction. PR-AUC is a fairer assessment of performance. In some cases, even standard measures can lead to selecting a more biased classifier because the effect of connectance is strong. The amount of correction to apply to the training dataset depends on network connectance, on the measure to be optimized, and only weakly on the classifier.4. These results reveal that training machines to predict networks is a challenging task, and that in virtually all cases, the composition of the training set needs to be experimented on before performing the actual training. We discuss these consequences in the context of the low volume of data.

Blockchain in Supermarkets: Mitigating the Problem of Organic Waste Generation

This work presents a proposal for a solution to the specific problem of organic waste generated by supermarkets and understood as t merchandise of organic and perishable composition that could not be marketed during its validity period. The goal of this research is to propose a solution based on Blockchain technology in Chile, which would allow an immutable, decentralized, and validated transaction record to be kept. Such a record would enable supermarkets to trace the life cycle of those products that make up organic and perishable merchandise in a transparent, reliable, and scalable way. To this end, the problem is modeled using the Blockchain Hyperledger Fabric platform (an open-source platform started by the Linux Foundation), which is fed with relevant information and data on the status of a representative set of organic merchandise products. At the same time, a qualitative approach is proposed to gather the opinions of executives and logistics operators through semi-structured interviews, and considering a convenience sample. With a sample of 6 executives, it is understood how the proposal is perceived and its applicability in supermarkets and distributors. The data show that both obtaining information and making decisions about it are achieved in a distributed and collaborative way, allowing for reliable and agile traceability, thereby mitigating the low quality of the information provided by the actors that make up the supply chain. This service is perceived as desirable by both customers and operators. The model enables not only horizontal communications between suppliers, distributors, and consumers, but also vertical ones, and thus, ultimately, makes the company's income statement more efficient.

Trainable generator of educational content

<p>As the main problem of the research, the possibility of creating a universal educational platform that combines the possibilities of an online generation of educational content with the interface of the training process itself was considered. The methodology of the educational platform has been developed, in which the mass generation of content is carried out at random, based on simulation models of educational objects. A matrix interface is used, which allows performing custom operations by entering a sequence of typical operators. The system forms a reference base of operators, replenishing it from user solutions, which makes it possible to train and improve the system in order to provide methodological support to student users. An active demo layout of an educational content generator was created and tested, using the example of a specific problem from school mathematics. All methodological options function in the layout. There are three interface options: administrative, training and control. It was concluded that the approach based on the simulation of educational objects makes it possible to create a unified algorithmic platform that combines the functions of content generation with educational training. The system contains a unique option to teach yourself based on its interaction with students.</p>

The Revised Attentional Fixation on Suicide Experiences Questionnaire (AFSEQ-R) and its Relationship with Suicidal Ideation and Behavior

Introduction: Attentional fixation is a cognitive process characterized by a preoccupation with a certain topic. The Attentional Fixation on Suicide Experiences Questionnaire (AFSEQ) was developed to measure fixation on suicide. The current study aimed to investigate a revised version of this measure (AFSEQ-R) and examine differences between recent suicide attempters and those with suicidal ideation (SI) only. Method: Participants were 57 inpatients who attempted suicide within 14 days of study participation and 57 inpatients who presented with SI but no suicidal behavior within the past year. Analyses included an exploratory factor analysis (EFA) of the AFSEQ-R, bivariate correlations between AFSEQ-R total score (and its factors) with related constructs, and logistic regression of attentional fixation as a moderator of the association between SI and attempt status. Results: EFA revealed a two-factor structure for the AFSEQ-R: Cognitive Stuckness (and AFSEQ-R total score) correlated with SI, anxiety, impulsivity, and specific problem-solving deficits, while Cognitive Dysfunction acted as a moderator of the association between Stuckness and SI. AFSEQ-R scores did not moderate the relationship between current SI and attempt status. Conclusion: Attentional fixation on suicide may contribute to the distress associated with SI. Interventions aimed at reducing attentional fixation may reduce suffering.

Interpolation on the special orthogonal group with high-dimensional Kuramoto model

The construction of smooth interpolation trajectories in different non-Euclidean spaces finds application in robotics, computer graphics, and many other engineering fields. This paper proposes a method for generating interpolation trajectories on the special orthogonal group SO(3), called the rotation group. Our method is based on a high-dimensional generalization of the Kuramoto model which is a well-known mathematical description of self-organization in large populations of coupled oscillators. We present the method through several simulations and visualize each simulation as trajectories on unit spheres S2. In addition, we applied our method to the specific problem of object rotation interpolation.

Iodine doping of a photocuring, one-component embedding medium for µCT and histology

Objective: Evaluation of μCT scans of bone implant complexes often shows a specific problem: if an implant material has a very similar radiopacity as the embedding medium (e.g. methacrylate resin), the implant is not visible in the μCT image. Segmentation is not possible, and especially osseointegration as one of the most important parameter for biocompatibility is not evaluable. Methods: To ensure μCT visualisation and contrast enhancement of the evaluated materials, the embedding medium Technovit® VLC7200 was doped with an iodine monomer for higher radiopacity in different concentrations and tested regarding to handling, polymerisation, and histological preparation, and visualisation in µCT. Six different µCT devices were used and compared with regard to scan conditions, contrast, artefacts, image noise, and spatial resolution for the evaluation of the bone-implant blocks. Results: Visualisation and evaluation of all target structures showed very good results in all μCT scans as well as in histology and histological staining, without negative effects caused by iodine doping. Subsequent evaluation of explants of in vivo experiments without losing important information was possible with iodine doped embedding medium. Conclusion: Visualisation of implants with a similar radiopacity as the embedding medium could be considerably improved. µCT scan settings should be selected with the highest possible resolution, and different implant materials should be scanned individually for optimal segmentation. µCT devices with higher resolutions should be preferred. Advances in knowledge: Iodine doped embedding medium is a useful option to increase radiopacity for better visualisation and evaluation of special target structures in µCT.

Multiple Region Coverage Path Planning for Autonomous Underwater Vehicle.

Coverage path planning methodology for an autonomous underwater vehicle to search multiple non-overlapping regions has been proposed in the paper. The proposed methodology is based on the genetic algorithm (GA). The GA used in the proposed methodology has been tuned for the specific problem, using design of experiment on an equivalent travelling salesman problem benchmark instance. Optimality of the generated paths was analysed through simulation studies. Results indicated that the proposed methodology generated shorter paths in comparison to conventional methods.

Game Mechanics: Sniper Simulator Game

Popularity of game creates opportunity to everyone in developing a game that can be used to address any specific problem in many areas. In BINUS University International, for education purposes, game is not only being used as a tool/media to learn something, but it can also be used to develop student understanding on how the game mechanic of the games works. By understanding the mechanic of the games, hopefully it can help them in developing a game designed that can be used to address problems in their environment.